Re: [Xen-users] Announcing XenMaster

[Jeff Sturm <jeff.sturm@xxxxxxxxxx>]

I'm reluctant to further a thread that's already off-topic, but two of these 
sub-topics cause me considerable heartburn...

> -----Original Message-----
> From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx [mailto:xen-users-
> bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Linus van Geuns
> Sent: Sunday, February 12, 2012 7:44 AM
> 
> > To conclude: Java works (for us). One can only hope its reputation gets 
> > better in the
> years to come.
> 
> Actually, besides being somewhat sluggish and fat ..and pretty verbose in its 
> language
> part, Java had a pretty good reputation.

"reputation" is the operative word in your post.  Ii may or may not be 
supported by facts.  Words like sluggish/fat/verbose are merely opinions 
without supporting arguments (and I don't want to argue them here).

The choice of development language is the wrong debate anyway.  There are 
plenty of examples of high-quality code written in Java, and you'll generally 
find these are simple to install, easy to run, and perform well.  I've also 
been forced to use Java code that produces frequent OOM errors, NPE exceptions, 
deadlocks, etc.  But I doubt it's difficult to find good and bad examples of 
just about any development language.

It may be that the average experience level of Java programmers is shorter than 
for other languages, as someone else suggested.  That may be important to 
understanding Java's reputation, but it's not important to the XenMaster 
project.  I personally welcome the opportunity to work with another well 
engineered, carefully designed OSS project regardless of choice of development 
language.

BTW, please don't remind me what Oracle is or isn't doing with Java.  I've 
heard it before.  Leave that discussion to the tabloids, not a technical user 
forum.

> @XenMaster: I would have expected any current Xen management front-
> end/framework to make use of and help advaince the libvirt project.

Depends on the project's goals.  My understanding of libvirt is that it is 
meant to be an abstraction layer to support Xen, KVM and other hypervisors.  
That's not important if the project's sole focus is Xen/XCP.  Libvirt won't add 
anything in that case, and it may in fact get in the way.  (My own, brief 
experiments with libvirt involved installing packages, following the 
documentation, troubleshooting libvirt and ultimately ignoring it.  Then 
everything began to work.  It was clear the Xen packages I was using at the 
time were more mature and/or fully tested than libvirt.  YMMV.)

-Jeff



_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users